On Mon, 13 Sep 2004 23:27:34 +0200 Joerg Heinicke <[EMAIL PROTECTED]> wrote:
> It seems to be heavy to do this transformation multiple times though > you need to do it only once using the FragmentExtractorTransformer. > Could help to reduce the memory needs. Well, after looking for some docs, it's of course the most elegant method, but less documented for this beautiful solution. >From reading the source here a example for Lincoln: In the sitemap or subsitemap: inside the map:component-section/map:transformars <map:transformer name="wordimgextractor" src="org.apache.cocoon.transformation.FragmentExtractorTransformer" logger="sitemap.transformer.extractor"> <extract-uri>http://schemas.microsoft.com/office/word/2003/wordml</extr act-uri> <extract-element>pict</extract-element> </map:transformer> the wordml2html-pipeline <map:match pattern="....."> <map:generate src="..."/> <map:transform type="wordimgextractor"/> <map:transform src="wordml2html"/> <map:serialize type="html"/> </map:match> This will remove all w:pict-elements from the wordml-document and store it inside the cocoon-cache and instead of the pict-elements you will get something like: <fe:fragment xmlns:fe="http://apache.org/cocoon/fragmentextractor/2.0" fragment-id="8b5dcbaa795d39f5"/> with this fragment-id you can query (as ImageID for the next pipeline) the stored image-data like <map:match pattern="*.jpg"> <map:generate type="extractor" src="{1}"/> <map:transform src="stylesheets/wordml2svg.xsl"> <map:parameter name="image" value="{1}"/> </map:transform> <map:serialize type="svg2jpeg"/> </map:match> where you get the w:pict-element back from the cocoon-cache and only this, so there is no more the whole wordml-document as DOM in the memory (with all the other images) like Joerg pointed out. In any cases (448MB or 2GB) this will spare your(Lincoln's) memory. Best Regards, Simon --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]